On Thu, May 21, 2015 at 12:20:06AM +0000, Kuninori Morimoto wrote:
Hi Mark
Thank you for your reply
Here, sound A/B/C will be used randomly. Does ALSA SoC already has this kind of system ?
I *think* at least the Intel systems do this already.
Thank you. I will check it.
I would like to ask you how to mix multi input in ASoC. Renesas sound device(s) can mix multi input to 1 output.
You mean in DPCM rather than a plain CODEC?
(snip)
codec will be started if one of A/B/C are started, and it will be stopped if all of A/B/C are stopped.
If you connect multiple front ends to one back end this should work (with DAPM routes and so on). If it doesn't work then that's something that should work.
I considered about DPCM, but *maybe* this is not goot match for us. If my understanding is correct, and if I used DPCM for this, all of sound A/B/C are start/stop:ed in same time ?
ex) start sound => start A/B/C and codec stop sound => stop A/B/C and codec
I guess it works if it has this connection
data -> [A] -> [B] -> [C] -> codec
We would like to do is like this
+-+dataA -> [A] -> |M| dataB -> [B] -> |I| -> codec dataC -> [C] -> |X| +-+
init => working: start sound A => working: codec + A start sound B => working: codec + A/B (mixed) stop sound A => working: codec + B start sound C => working: codec + B/C (mixed) stop sound B => working: codec + C stop sound C => working:
I guess current DPCM doesn't support this ?
it does, you have to declare A, B, C as FE then introduce a digital mixer (real or virtual, depending on the system) and the a BE which is your HW DAI here and connects to codec
This works very well and you can start/stop A, B, C independently. Codec will be on as along as any of the inputs is On.
See the sound/soc/intel/atom for reference